"
x2Real wrote:
"
f3rret wrote:
Oh hey, it's another person expecting GGG to fix stuff that is partially unfixable...
Furthermore, it's also a person who thinks that overhauling the code GGG has spent *years* creating will be a quick process.
maybe if they weren't incompetent in the first place
oh look! another idiot who knows nothing he's talking about.
/forum/view-thread/869543 all my builds' compilation
Please PM me on forum to inform me if I won an auction. Thank you.
|
Posted byabcnever#3020on Apr 5, 2014, 5:58:32 PM
|
"
f3rret wrote:
meaning one point has to handle all the work for the server.
Just no. CBA to explain to you why. How do you think the serversystem behind MMOs like WoW work?
"
Gobla wrote:
"
ReZar wrote:
Except you are wrong, I've seen many desync threads in my time here and I've seen no proof that it can be fixed and I really won't take your word for it - it can obviously be made better than it is but it can't be fixed.
I know some people without any knowledge have made threads about making predictions client side and such but that is not a fix since it will make the game more open to exploits.
1. Predictions client-side do not open any exploits at all. If they opened up exploits they wouldn't be predictions. That's why it's a prediction, it's predicting what's happening server-side, it's not determining what's happening server-side. A prediction can not, by definition, open up an exploit.
2. An entirely different networking model would fix it while having equal protections against exploits and hacks as the current model. A model that waits for the server, used in pretty much every MOBA which often feature many of the same mechanics (including chance to hit, stuns, evasion, block etc.) PoE has. Now this model has other downsides (quality of play reduces greatly with higher pings) but the sad part is that each of those downsides is also present in PoE frankly horrible implementation of the client-side prediction model. PoE also becomes very much a pain to play with pings above 200, just like the alternative.
Desync itself can't be fixed in the current networking model, there are alternatives that are widely used in other games that don't suffer from desync but do suffer from high pings.
PoE's implementation of this networking model unfortunately suffers from both desync AND high pings. It's literally the worst of both worlds right now.
This would be one possibility. Another would be to just sync more fucking often. There are some incredibly easy fixes which would increase QoL by a lot. Like syncing whenever a client issues an order which can't be executed without moving (just add a fucking bit to the message stating if the client thinks he needs to move to execute too).
|
Posted bywlfbck#6588on Apr 5, 2014, 6:17:24 PMBanned
|
|
Its kinda funny how people always say the problems can't be fixed because its inherent. In Dota2 there is 0 unit delay and if you lag out everything freezes on your screen momentarily and resyncs you when you reconnect.
In Diablo 3 they've basically fixed the entire desync problem.
|
Posted byCrnm#5855on Apr 5, 2014, 6:27:29 PM
|
"
Crnm wrote:
Its kinda funny how people always say the problems can't be fixed because its inherent. In Dota2 there is 0 unit delay and if you lag out everything freezes on your screen momentarily and resyncs you when you reconnect.
In Diablo 3 they've basically fixed the entire desync problem.
They're called "Desync Apologists" and the majority of them also have multiple "pretty titles" next to their name.
Coincidence?
Nah... More like, they're trying to justify it, to themselves.
|
Posted byStinkFinger#4623on Apr 5, 2014, 6:29:57 PM
|
"
Crnm wrote:
Its kinda funny how people always say the problems can't be fixed because its inherent. In Dota2 there is 0 unit delay and if you lag out everything freezes on your screen momentarily and resyncs you when you reconnect.
In Diablo 3 they've basically fixed the entire desync problem.
Well it can't be "fixed" in the very direct meaning of the word. But it can be worked around enough so that it's not a concern anymore with a normal ping.
But the desync apologists just want to be right so they get super-anal and say "it can't be fixed" meaning the very direct meaing, knowing full well that everyone is talking about reducing it's effect so it isn't a problem anymore.
(Also i'm not exactly sure why anyone would defend this shit, there is nothing to be gained *at all* by doing that)
|
Posted bywlfbck#6588on Apr 5, 2014, 6:31:15 PMBanned
|
|
Is desync worth shutting down all other development for months? And then delaying all future development, so that each update will be fully optimized?
No, it's not. GGG will just keep pushing out more content to appease the people who can deal with the desync. They can't afford to write and rewrite the code until they get it just right.
Hopefully they can make their current setup deal with entropy, else the buildup will eventually make things worse all around.
Shmo's Suggestions; Dozen Skills!
http://www.pathofexile.com/forum/view-thread/402487
|
Posted bywdmshmo#3292on Apr 5, 2014, 6:45:55 PM
|
"
wdmshmo wrote:
Is desync worth shutting down all other development for months? And then delaying all future development, so that each update will be fully optimized?
Well, YMMV. I think it is, considering that 90% of my deaths go to it. and having to /oos all the time isn't exactly fun either.
|
Posted bywlfbck#6588on Apr 5, 2014, 6:52:23 PMBanned
|
"
wdmshmo wrote:
Is desync worth shutting down all other development for months? And then delaying all future development, so that each update will be fully optimized?
Why would they need to shut down other development? I keep hearing the excuse that the artist guy cant fix desync, so while desync is being fixed by whoever, the artist can normaly work on content.
Just using some inverse whiteknight logic :P
E:
But serious answer: I wouldn't have a problem with it. Can easily deal with a shutdown or months long stagnation, if it would mean that in half a year there would be no or minimal desync (and optimized graphics too). Would come back to play the game for sure.
But yeah, it wouldnt be a good business decision in general.
When night falls
She cloaks the world
In impenetrable darkness Last edited by morbo#1824 on Apr 5, 2014, 7:22:14 PM
|
Posted bymorbo#1824on Apr 5, 2014, 7:09:48 PM
|
"
CharanJaydemyr wrote:
"
wlfbck wrote:
This would be one possibility. Another would be to just sync more fucking often.
loong text
I'm aware of the stuff you wrote, but there is still a lot of middleground they could cover. I'm well aware that syncing all the time results in an unplayable mess, but i'm talking about these situation where one actually wanders off 10-15meters to get to an enemy. This can take up to several seconds. A possible solution for example would be: not letting the player move, sync back mob position to client (and possibly player position as well if it's really far off) and having the player attack in the general direction of the mob, even if it hits a wall (chances are big there is more than one monster).
Also, they could just give a little leeway to the client. There really isn't much realistic potential for a cheat evolving from that.
But i already wrote all this in the past, it got ignored, as always. I'm just fed up with it and am just play (semi-)caster aoe builds now with holding shift all the time. It's less annoying that way.
|
Posted bywlfbck#6588on Apr 5, 2014, 7:19:53 PMBanned
|
"
CharanJaydemyr wrote:
"
wlfbck wrote:
I'm aware of the stuff you wrote, but there is still a lot of middleground they could cover. I'm well aware that syncing all the time results in an unplayable mess, but i'm talking about these situation where one actually wanders off 10-15meters to get to an enemy. This can take up to several seconds. A possible solution for example would be: not letting the player move, sync back mob position to client (and possibly player position as well if it's really far off) and having the player attack in the general direction of the mob, even if it hits a wall (chances are big there is more than one monster).
Also, they could just give a little leeway to the client. There really isn't much realistic potential for a cheat evolving from that.
But i already wrote all this in the past, it got ignored, as always. I'm just fed up with it and am just play (semi-)caster aoe builds now with holding shift all the time. It's less annoying that way.
Dude, you're asking them to fundamentally change the client-server relationship with all those solutions. Not. Happening.
In sequence:
Not letting the player move: the backlash from players who *don't* read these forums and suddenly don't move when they click would be disastrous
resyncing more often/specifically: I covered that
attack in the general direction of the mob: Not quite understanding that one.
Leeway to the client: you...didn't really read what I wrote, did you? There is no leeway. The server must control everything! The Client is a whore! NO TRUST! NONE!
And yes, holding shift is integral. Forces you to attack what's in front of you rather than what you *think* is in front of you. Is that what you meant by attack the general direction of the mob btw?
Yeah that's what i meant with "attack in general direction" (could've worded that better :P ).
I know they will never give some leeway to the client :(
The other stuff basically is very implementation specific and something that would need actual testing to judge how effective/bearable it is. Which comes down to what you first said in this quote: not. happening. (very unfortunate)
|
Posted bywlfbck#6588on Apr 5, 2014, 7:30:39 PMBanned
|
